Interaction of Sox1, Sox2, Sox3 and Oct4 during primary neurogenesis

Abstract

Sox1, Sox2 and Sox3, the three members of the SoxB1 subgroup of transcription factors, have similar sequences, expression patterns and overexpression phenotypes. Thus, it has been suggested that they have redundant roles in the maintenance of neural stem cells in development. However, the long-term effect of overexpression or their function in combination with their putative co-factor Oct4 has not been tested. Here, we show that overexpression of sox1, sox2, sox3 or oct91, the Xenopus homologue of Oct4, results in the same phenotype: an expanded neural plate at the expense of epidermis and delayed neurogenesis. However, each of these proteins induced a unique profile of neural markers and the combination of Oct91 with each SoxB1 protein had different effects, as did continuous misexpression of the proteins. Overexpression studies indicate that Oct91 preferentially cooperates with Sox2 to maintain neural progenitor marker expression, while knockdown of Oct91 inhibits neural induction driven by either Sox2 or Sox3. Continuous expression of Sox1 and Sox2 in transgenic embryos represses neuron differentiation and inhibits anterior development while increasing cell proliferation. Constitutively active Sox3, however, leads to increased apoptosis suggesting that it functions as a tumor suppressor. While the SoxB1s have overlapping functions, they are not strictly redundant as they induce different sets of genes and are likely to partner with different proteins to maintain progenitor identity.

Fig. 1. Oct91 and Sox1 expand the neural tube and repress epidermal formation similarly to Sox2 and 3

(A-D) Bright field and fluorescent microscopy of stage 17 embryos injected with Sox1, Sox2, Sox3 or Oct91 and GFP mRNA into one cell of a two-cell stage embryo. (E-H) WISH for epi-k expression in embryos injected with soxB1 and LacZ, or oct91 and GFP mRNA. Asterisk in E marks the injected side. The injected sides of Sox1 (180/190), Sox2 (135/152), Sox3 (81/81) and Oct91 (49/53) have repressed epi-k expression compared to their uninjected side. All views are dorsal with anterior to the left. Brackets in A-D compare the relative width of the neural tube of the uninjected (top) to the injected side (bottom).

Fig. 2. Oct91 and the SoxB1s synergize to induce posterior protrusions

(A-L) WISH for n-tub in embryos injected in one blastomere of the four-cell stage embryo with LacZ and oct91, sox1, sox2, or sox3 mRNA. Arrowhead in B marks laterally displaced neurons. Ectopic neurons are seen at both stage 23 and 30 in all treatments. (M-V) WISH for n-tub in embryos injected in one ventral blastomere of a four-cell embryo with oct91 alone or in combination with sox1, sox2 or sox3, and LacZ. Asterisk in top row marks the injected side. In A-H and M-R the injected side is the bottom half of the embryo. In I-L and S-V only the injected side is shown.

Fig. 3. Sox3 induces neural gene expression in gastrula-staged ectodermal explants

(A) Diagram of ectodermal explant assay. mRNA is injected into both blastomeres of the two-cell embryo. Explants are excised at stage 8-9. Cultured explants that are ubiquitous for GFP are collected and processed by RT-PCR. (B) RT-PCR analysis of stage 12 ectodermal explants that are uninjected (UI) or injected with noggin, sox2 or sox3, sox2 plus sox3. A whole embryo (WE) control is also shown. In the co-injected caps, the total amount of RNA injected equals that of sox2 or sox3 injected singly. The caps are negative for bra and therefore do not have mesoderm contamination. (C) Diagrams of Sox3 and Sox3-VP16. Twenty amino acids were removed from the C-terminus and replaced with VP16 (Diagram adapted from (Kamachi et al., 2000).

Fig. 4. Sox2, Sox3 and Oct91 induce unique arrays of neural genes in ectodermal explants

(A) RT-PCR analysis of stage 22 whole embryo (WE) or ectodermal explants from embryos injected with noggin (Nog), sox2 or sox3, sox2 (S2) with sox3 (S3), sox3-VP16 or uninjected (UI). (B) RT-PCR analysis of stage 22 ectodermal explants from embryos injected with oct91, or a low doses of oct91 with sox2 or sox3.

Fig. 5. Sox1 induces early neural markers and counteracts Oct91-mediated neuron formation

RT-PCR analysis of stage 22 embryos (WE) or ectodermal explants from uninjected (UI) embryos, embryos injected with noggin (Nog), sox1 low, oct91 high, or oct91 plus sox1 (S1). GFP was injected in all treatments as a lineage tracer.

Fig. 6. Oct91 knockdown depresses neural induction by Sox2 and Sox3

(A) RT-PCR analysis of uninjected stage 12.5, 16 and 22 whole embryos (WE) or ectodermal explants. (B) RT-PCR of ectodermal explants that are either uninjected (UI), or injected with noggin (Nog), sox1, sox2, or sox3 alone or in combination (+) with Oct91 morpholino (MO). The caps are negative for m act and therefore do not have mesoderm contamination.

Fig. 7. A constitutively-active bicistronic vector generates two proteins from a single transcript in Xenopus

(A) A diagram for the bicistronic 2A vector shows that a single transcript forms two proteins. mCherry (pink) contains a plasma membrane localization signal (PmLS, white) and GFP contains nuclear localization signal (NLS, black). (B-E) Xenopus embryonic fibroblast cell transfected with the bicistronic system with vector in A. mCherry is in the plasma membrane (C) and GFP in the nucleus (D). (F-H) In vitro transcribed 2A vector mRNA injected into half of the two-cell stage embryo is visible under a dissecting fluorescent microscope in a neurula stage 17 embryo. (I) Bicistronic constructs used for the generation of transgenic embryos. Each SoxB1 (purple) is upstream of the 18 amino acid 2A viral sequence (blue) followed by nuclear localizing GFP (green) and poly-A signal (brown) and expression is controlled by the constitutively-active EFlα promoter (orange). Asterisk marks injected side. Diagram not to scale.

Fig. 8. Sox1:2A:GFP and Sox2:2A:GFP transgenic embryos have decreased levels of n-tub and Sox2:2A:GFP have increased proliferation

(A-C) WISH for n-tub. (A) A stage 30 sibling embryo generated from sperm nuclei transfer serves as a staging control. N-tub levels are reduced in the brain and spinal cord of transgenic embryos. (B) Sox1:2A:GFP and (C) Sox2:2A:GFP transgenic embryos. (D-F) WISH for sox2 in stage 30 transgenic Sox1:2A:GFP embryos. Sox2 is not expanded in embryos with small heads or absent heads. (G-I) Immunohistochemistry for phosphorylated histone H3 immunohistochemistry in Sox2:2A:GFP transgenic embryos. Embryos are shown at the same magnification. Hatched box indicates magnified view. (G’-I’) pH3 positive cells have been counted in the magnified views.

Fig. 9. Continuous expression of sox3 increases cell death

TUNEL staining in (A) uninjected embryos or those injected with (B, E) Sox2:2A:GFP or (C, F) Sox3:2A:GFP plasmid DNA at the 1 cell stage. (A-C) Stage 11 embryos are animal pole views. (D-F) Stage 20 embryos are lateral views with anterior to the left. Numbers of embryos with the represented phenotype are in the lower right corner. (G) Percentages of embryos with relatively high (more than 50 cells positive), low or no staining.

Fig. 10. Summary of genes induced and cell fates altered by each SoxB1

Overexpression of each SoxB1 in ectodermal cells inhibits epidermis formation, which is driven by BMP signaling. Sox3 overexpression leads to the formation of neural progenitors by inducing expression of gem, msi1, eomes, sox1, and FGF8. Sox1 and FGF8 are gray because their induction is not maintained past stage 12. For neuronal differentiation, Sox3 requires the presence of a co-factor. Sox2 induces expression of sox1 and eomes in explants. Sox1 induces expression of both sox2 and sox3 and ultimately results in differentiation and the formation of neurons.